Enhancement of Online Stochastic Gradient Descent using Backward Queried Images
Stochastic gradient descent (SGD) is one of the preferred online optimization algorithms. However, one of its major drawbacks is its predisposition to forgetting previous data when optimizing through a data stream, also known as catastrophic interference. In this project, we attempt to mitigate this drawback by proposing a new low-cost approach which incorporates backward queried images with SGD during online training. Under this new approach, we propose that for every new training sample through the data stream, the neural network is optimized using the corresponding backward queried image from the initial dataset. After compiling the accuracy of the proposed method and SGD under a data-stream of 50,000 training cases with 10,000 test cases and comparing our algorithm to SGD, we see substantial improvements in the performance of the neural network with two different MNIST datasets (Fashion and Kuzushiji), classifying the MNIST datasets at a high accuracy for the mean, minimum, lower quartile, median, and upper quartile, while maintaining lower standard deviation in performance, demonstrating that our proposed algorithm can be a potential alternative to online SGD.
A New Method For Microplastic Removal and Optical Measurement
Microplastics are tiny invisible plastic pieces that are piling up in the marine environment emerging as one of the many environmental issues which our planet is facing today. Researches for the removal of these particles are important because studies that have been made so far haven't come up with an effective solution. This project aimed to detect microplastics and remove them from aqueous environments with an effective and practical method then it was aimed to determine the removal amount of microplastics by optical measurements with the developed system. Firstly, the magnetic carbonanotubes (m-CNT) which is intended to hold onto the surfaces of microplastics was synthesized and added to the mixture of microplastics. Then the magnet within a glass tube was passed through the mixture and the sample was cleared of microplastics. A spectrometer was made to monitor this process and after its calibration, it was used to measure coffees with different concentrations. It has been shown that their concentrations can be determined by calculating the transmission values and Rayleigh scattering. In the end, it has shown that there are no micro or nano-sized plastic particles when removed with M-CNT, within the accountable range of the spectrometer that had been made. Hence the removal of the microplastics: an invisible threat for the environment has been studied by combining nanomaterials with unique surface properties in the removal process and an optical principle such as Rayleigh scattering, a new technique has been developed that can measure quickly, economically,